Spectrum analysis of radiotracer residence time distribution for industrial and environmental applications

Radiotracer signal analysis and recognition still represents challenges in industrial and environmental applications specially in residence time distribution (RTD) measurement. This paper presents a development for the RTD signal recognition method that is based on power density spectrum (PDS). In this development, the features are extracted from the signals and/or from their higher-orders statistics (HOS) (Bispectrum and Trispectrum) instead of PDS. The HOS are estimated using direct, indirect and parametric estimations. The recognition results are analyzed and compared for different HOS estimation in order to select the best HOS estimation method for the purpose of RTD signal recognition. The artificial neural networks are used for training and testing of the proposed method. The proposed method is tested using RTD signals obtained from the measurements carried out using radiotracer technique. The simulation results show that the parametric estimation of the Trispectrum gives the higher recognition rate and is the most reliable for the RTD signal recognition.